Tribute to the GEEKS of this Forum.

[uasish]

Are some people trying to Re-Invent Wheel ?
.........................................................................
Of late few terms like 'time series' / 'data mining' / 'Random' / etc etc are being
related to some thing of irrelevance than Basic indicators ( RSI / MACD etc),what
we learnt & found time & again in your experience to work in actual Mkt situation.
Hence those Guys are subtlying trying to propaganda their 'Knowledge'
they are at best 'Masterji' but drifting away from actual objective of Wealth Creation. These 'Geeks' objective is for PHD than money earning.
This post is of my personal inferences,on above subjects.
What is RSI ?
In layman's term ,in June 1978 ,in an article of Futures magazine Welles Wilder,
devised this Oscilator with a fairly simple formula
U= An Avg of upward price change,
D= An Avg of downward price change,
100 - [ 100 / ( 1 + (U/D) ] this is normally done on Close ,with different time period.
We will refer this at later stage of this post.
What is Time series ?
In statistics, signal processing, and econometrics, a time series is a sequence of data points, measured typically at successive times, spaced at (often uniform) time intervals. Time series analysis comprises methods that attempt to understand such time series, often either to understand the underlying theory of the data points (where did they come from? what generated them?), or to make forecasts (predictions). Time series prediction is the use of a model to predict future events based on known past events: to predict future data points before they are measured. The standard example is the closing price of a share of stock based on its past performance.
So now all CLOSE on the right hand side of chart (will come on Monday) at specific interval (say EOD), CAN THIS BE A TIME SERIES ? Yes no confussion this closes or / O / H / L etc are variables (becoz everyday it varies ,not constant) in respect to TIME (each end-of-day),hence nothing but a Time Series Data.
So why these geeks are confusing us ? by not mentioning close but with a swanky word like 'Time Series' ? Becoz by renaming they suddenly Hit us in
conceptual level,broadening the horizon.How ? Let us dissociate from chart CLOSE & shift to say Appolo Hospital's Intensive Coronary Care Unit,a heart patient ,on constant monitering programme,his ,mine (incidentally Ratan's heart in remembrance of Vabhna)all beats per minute ,means each beat=intraday data ,beats per min=close,every min=EOD,is also a Time Series.If the patient has Tacchy cardia(above normal beats/ min) or Bradycardia(below normal beats/min)
These signals of heart ,means heart's time series is transfered by ECG to a chartical format,which the Cardiologist can read,he can actually Predict a below normal beat ( a selling phase in say Bajaj Hindusthan) to be a partial RBB(right bundle block) his terminology for RSI's divergence/convergence,& he convey's to patient party nothing to worry about,this is actually a manifestation of Diabetic Neruropathy of the patient(forget terminologies of the Dr read it ,a Financial Analysts reading of Sugar sector vis a vis county's GDP ,blah blah) .
What is Montek Singh Aluwalia at Planning Commission & RBI governor Dr Reddy doing day in & day out ? Grasping this Time Series to Forecast.The GM of Volkswagon or Nandan Nilkeni of Infosys is giving guidance ,means forecasting by analysizing time series data.
Hence all the World's knowledge of forecasting based on Time Series Data is now at our doorstep,their methods of tackling Random ,India's vast terrain with it's imensely scattered economic activities when its GDP can be forecasted accurately (except to the decimal points),why dont we explore the Modules they are using ,so as to forecast the right side of the chart ?Let us plz take advantage of this Huge Knowledge base to better forecast To-morrow.This conceptual change was the purpose of these PHD wallahs.
Now to Data & approach to data bank.

Data Mining
Data Mining is an analytic process designed to explore data (usually large amounts of data - typically business or market related) in search of consistent patterns and/or systematic relationships between variables, and then to validate the findings by applying the detected patterns to new subsets of data. The ultimate goal of data mining is prediction - and predictive data mining is the most common type of data mining and one that has the most direct business applications. The process of data mining consists of three stages: (1) the initial exploration, (2) model building or pattern identification with validation/verification, and (3) deployment (i.e., the application of the model to new data in order to generate predictions). (CV's post has opened my eyes)
In layman's term it is a prefabricated data,before we do any analysis on it.
We all know RSI(14)
Tell me why we normally use close ,have we used Median Price instead of close.
For any Large candle close is more skewed ,becoz of the High Range of that bar.
In case of Divergence/Convergence try Median Price a more clear picture will emerge. This Median price use is one kind of crude data mining.
Now let us revert back to W.Wilder's formula of RSI(14) he has mined (prefabricated) in 2 variables U & D of the Time Series (close) in a sample test (drawn out 14 periods from say 10000 bars) there can be another method of sampling of data,unlike last 14 days ,we can at random take representations of this last 10000 bars & with a reference point say yesterday's close (substract it to get Up & Down & those difference figs ,convert to log scale & then make a RSI .
There can be 100 methods but the logic for doing that has to come from my conceptual understanding & why ? all these are Data Mining & Time Series Forcast.These guys are actually helping us ,to discard their concept will be like my attitude towards computer few years back,which will retard my pace to "Wealth Creation"
Thks to all senior members who have actually by their posts helped me & helping me.

 

[uasish]

http://www.traderji.com/88389-post932.html

It jolted me for Data being Non- Stationary :

I reasoned with my understanding but later on now to-day agree with him;
Excerpt from yesterday late night SavantGarde's mail
Non-stationary time series. Non-stationarity means that statistical properties of the signal such as mean, standard deviation are not conserved as function of time. For example, if we measure the average of the first N points and of the whole signal, the values we will get won't be the same. DFA (Detrended Fluctuation Analysis algorithm )should handle with the nonstationarities in signal by determining local polynomial trends and observe the fluctuations around these trends.
So first we should integrate our signal to convert it to a self similar process. Usually we also detrend the signal (remove the global average) before doing integration. This step is necessary only for preventing the overflow of numerical calculations when integrating long data series with non zero average.
Before going straight to the details of the DFA algorithm, let's focus on some useful terms. Many of the real objects in the world satisfy so called fractalproperties : self-similarity and fractional dimension. Conventional Euclidean geometry is unable to describe nature in all the details . Fortunately, all the things around us don't look like spheres, pyramids or cylinders but have much richer and complex structure. All Euclidean geometrical figures have integer dimension and this fact limits our capacities to create new structures and describe the nature. However, fractal geometry was designed to work with objects of fractional dimensions, and that's why it describes the nature much better. Many objects possess also self-similarity properties, this means that if will zoom into some area of the object we will see that it looks remarkably similar to the whole object.
Hence CV is right ,when he says no need to study Indicators in isolation.My inference :
When the whole world's data bank use these 'Satistical' module to Forecast;
A cardiologist ,forecast's LIFE ,an economist forecast's the MOMENTUM & the possible Fate of Billions,a banker weighs & balances the OPTIMAL ,all of these noted distinguished scholars ,having access to immense resources at their command,decides on TIME series to Forecast THE RIGHT SIDE OF THE CHART.
Now are people's using this method profitably ?
Another excerpt from Net:
What if scientific methods and approaches are applied to stock markets? The towering figure of this approach is ace speculator Victor Niederhoffer, a legend on Wall Street. He has pioneered what is called the quantitative study of the market, which relies on a scientific approach. According to Niederhoffer, the starting point of this approach is defining the scientific method itself, which the Oxford English Dictionary defines as “a method of procedure that has characterised natural science since the 17th century, consisting in systematic observation, measurement and experimentation and the formulation, testing and modification of hypotheses.”
Applying mathematics to markets is not rare. Victor is only the most celebrated example. There are boutique research firms that apply high-end computers to crunch market data applying sophisticated mathematical tools. Victor trades strictly on the basis of statistical anomalies and the quantification of persistent psychological biases. He owes his influence to Francis Galton, the inventor of weather maps. Victor came across Galton while studying the influence of weather on sporting outcomes. Galton’s favourite motto was, “Wherever you can, count.” He traveled with a pin and an index card so that he could tabulate anything that struck him as noteworthy.” As Galton himself said it in his book Memories of My Life: “I frequently make statistical records of form and feature, in the streets or in company, without exciting attention, by means of a fine pricker and a piece of paper…The holes are easily counted at leisure, by holding the paper against the light, and any scrap of paper will serve the purpose...”
Historical data is easy to get and understand. Any research that goes beyond past performance numbers is a labour-intensive, time-consuming process and
demands experience and knowledge of what to look for.
This is not prank real life living characters.
"T"he advantage scientists bring into the game is less their mathematical or computational skills than their ability to think scientifically. They are less likely to accept an apparent winning strategy that might be a mere statistical fluke” - Jim Simons, President, Renaissance Technologies
Can somebody who holds seminars on topics like “Generalized Chern-Simons Invariants as a Generalized Lagrangian Field Theory” on his 60th birthday, be the best market player in the world?

In the last issue, I mentioned that there are at least two great investors (I use the term loosely to include traders), who have very successfully used quantitative tools to record stunning market performances over a long period. One of them is Jim Simons, about whom the Indian media is totally unaware. That is not their fault. Few even in Wall Street, including his fellow-travellers in quantitative trading, know the full dimensions of Renaissance Technologies Corp. that Simons heads.

Over more than two decades, Simons has been the leading light in marrying math and markets. Simons is supposed to be a cryptanalyst, mathematical physicist, academic - and a billionaire investor. Simons graduated from Massachussets Institute of Technology at 20 and completed his Ph.D. in Mathematics from the University of California at Berkeley, barely three years later. By 23, he was in the MIT faculty. From 1961 to 1964 he taught mathematics at MIT and Harvard University. In 1968, he became the chairman of Stony Brook’s math department and was beginning to be famous in academia. Simons did breakthrough discovery and application of certain geometric measurements that resulted in the Chern-Simons form (also known as Chern-Simons invariants, or Chern-Simons theory). In 1974, his theory was published in Characteristic Forms and Geometric Invariants, co-authored with the differential geometry expert Shiing-Shen Chern. The theory has a wide use in theoretical physics, particularly the famous string theory, used to explain the birth of the universe, black holes and supergravity.

In 1976, at 38, Simons won the American Mathematics Society’s Veblen Prize that is awarded every five years, for his work on differential geometry. It is the highest honour in the world of geometry.

Simons got interested in trading only in the early 1970s. In 1973, a tile company that he had invested in got sold and Simons gave his money to a mathematician who was trading in commodities. In eight months, Simons saw his money go up 10 times. Intrigued, Simons began trading in currencies. In 1978 he left academics, to form a private investment fund called Limroy. In 1982, he founded Renaissance Technologies Corporation, a private investment firm based in New York. Simons is still at the helm, as president, of what is probably the world’s most successful hedge fund.

In 1988 Simons decided to launch a fund that concentrated on pure trading. He launched Medallion in March 1988. The fund earned 8.8% in 1988 but lost money steadily thereafter until Simons stopped trading. For six months Simons and Princeton mathematician Henry Laufer (still Renaissance’s research chief), rebuilt Medallion’s trading strategy, shifting from fundamental analysis to a mathematical approach and rapid-fire trading strategy.

According to Institutional Investor magazine, Simons earned $1.5 billion in 2005 and $670 million in 2004. The only major article about him published by the Institutional Investor magazine in 2000, said that “Simons may very well be the best money manager on earth.” In March 1988 Simons started his flagship $3.3 billion Medallion fund. By 2000, the fund had notched returns of 35.6%, compared with 17.9% for the Standard & Poor’s 500 index. For 11 full years ending December 1999, Medallion’s cumulative returns were 2,478.6%. Quantum fund of George Soros was the next-best with a 1,710.1%. Want to invest in his fund?

Sorry, it has been closed to new investors since 1993.

So now tell me will it not be a Dumb act if i ignore these knowledge,Trading is my only source of (inconsistent )income for last so many years & will be in conceivable future.Hence the slight hint in a post here & there of these Geeks,which hits me conceptually ,i am trying to imbibe ,becoz i am here not to do a PHD but to earn,i have no pension ,no MIS in post office or FD in Bank all my assets are in my demat or my family members demat & all liquid cash in my a/c with the Broker's.Every month last week i make a payout,it is my salary.
My quest for Wealth Creation is my only Survival.
All RSI / MACD are price derivative ,means Mined Data,if those signals ,filtered with Satistical Module ,can give us a better Methodology to forecast,a better Win probability ? Hence let me try it,back test it.
One thing is an absolute certainity for a Systems Trader his/her
Systems is not going to work after sometime it will Peak Out,the Optimal f is going to reach an otimal level.Madam Linda Rashke ,it is being rumored, her excellent efficient Systems failing to perform like in old time.
My request to all senior members in this 'August Forum' ,plz guide me
so that i earn more,so that i can Retire comfortably.

 

[hitesh sabharwal]

you all have so much technical analysis knowledge,nowhere near you all.Amazing !!!

 

[Disciple]

Dear Asishda,
Very very good write up.It's really interesting to know about Jim Simons! You are such a knowledge hungry person! Waiting for more posts from you and also seniors of this forum.
God Bless You.

 

[ssprasad]

Ashish,
Very informative article ....

 

[ratan jain]

Uasish:

Nice article......

But I would suggest that u should not try to search for the holy grail
Its elusive, a mirage in the desert.......
Tell me, if u dont mind, whats ur winning % and ur winn/loss ratio? Maybe I can help ?

 

[uasish]

Ratan,
Thks for your concern for me,my win/loss ratio & winning % is somehow so so,i am not at all concerned about it ,rather nowadays trading has become stress free,i am worried about the global phenemenon of System Traders ,half life (time taken by a drug to be flushed out from the human system to kidney),all over ,the System which used to garner handsome profit has started loosing sheen,newer system are replacing older ones in % profit earning & avg profit per trade is diminishing.My endeavour is to keep the 'sharpe' high,& never let the weapon attaract rust,by continuous 'log'ing.
Stoped the 'Holy Grail' search when came to know from "da vincchi code" that it actually does not exist.
Asish

 

[beginner_av]

Are some people trying to Re-Invent Wheel ?
.........................................................................
Of late few terms like 'time series' / 'data mining' / 'Random' / etc etc are being
related to some thing of irrelevance than Basic indicators ( RSI / MACD etc),what
we learnt & found time & again in your experience to work in actual Mkt situation.
Hence those Guys are subtlying trying to propaganda their 'Knowledge'
they are at best 'Masterji' but drifting away from actual objective of Wealth Creation. These 'Geeks' objective is for PHD than money earning.

check this out

http://uk.reuters.com/article/stocksAndSharesNews/idUKNOA12716220070601?feedType=RSS

 

[ratan jain]

Hey no problem!
Part of article reproduced below

In 2006, an IBM Institute for Business Value study found that for every 40 traders that are active for a given product, there will be only four left by 2015 due to the electronification of markets.

"The four traders will be the stars that assume risk, achieve true client insight and, of course, consistently beat the market," according to the study, called "Tackling latency -- the algorithmic arms race."

And those survivors are set to do very well.

See, I told u I would prosper

 

[beginner_av]

Hey no problem!
Part of article reproduced below

In 2006, an IBM Institute for Business Value study found that for every 40 traders that are active for a given product, there will be only four left by 2015 due to the electronification of markets.

"The four traders will be the stars that assume risk, achieve true client insight and, of course, consistently beat the market," according to the study, called "Tackling latency -- the algorithmic arms race."

And those survivors are set to do very well.

See, I told u I would prosper

well if there are 2 million traders in India, 20000 will definitely be left and a lot more. thats all the members of traderji. so rejoice and if there are 10000 full time traders now, 1000 will be left. more full time traders than traderji has ever seen. again we are all there!!! the beauty of statistics!!!

do you remember the stats on Y2K problem?